Support tri-backquote style code block and fix ordered list indent (#431)

Co-authored-by: Alireza Savand <[email protected]>

Author: wetor

html2text/__init__.py

@@ -78,6 +78,7 @@ def __init__(
         self.use_automatic_links = config.USE_AUTOMATIC_LINKS  # covered in cli
         self.hide_strikethrough = False  # covered in cli
         self.mark_code = config.MARK_CODE
+        self.backquote_code_style = config.BACKQUOTE_CODE_STYLE
         self.wrap_list_items = config.WRAP_LIST_ITEMS  # covered in cli
         self.wrap_links = config.WRAP_LINKS  # covered in cli
         self.wrap_tables = config.WRAP_TABLES
@@ -111,6 +112,8 @@ def __init__(
         self.blockquote = 0
         self.pre = False
         self.startpre = False
+        self.pre_indent = ""
+        self.list_code_indent = ""
         self.code = False
         self.quote = False
         self.br_toggle = ""
@@ -629,6 +632,7 @@ def link_url(self: HTML2Text, link: str, title: str = "") -> None:
             self.lastWasList = False
 
         if tag == "li":
+            self.list_code_indent = ""
             self.pbr()
             if start:
                 if self.list:
@@ -644,15 +648,16 @@ def link_url(self: HTML2Text, link: str, title: str = "") -> None:
                     # TODO: line up <ol><li>s > 9 correctly.
                     parent_list = None
                     for list in self.list:
-                        self.o(
-                            "   " if parent_list == "ol" and list.name == "ul" else "  "
-                        )
+                        self.list_code_indent += "   " if parent_list == "ol" else "  "
                         parent_list = list.name
+                    self.o(self.list_code_indent)
 
                 if li.name == "ul":
+                    self.list_code_indent += "  "
                     self.o(self.ul_item_mark + " ")
                 elif li.name == "ol":
                     li.num += 1
+                    self.list_code_indent += "   "
                     self.o(str(li.num) + ". ")
                 self.start = True
 
@@ -715,8 +720,11 @@ def link_url(self: HTML2Text, link: str, title: str = "") -> None:
             if start:
                 self.startpre = True
                 self.pre = True
+                self.pre_indent = ""
             else:
                 self.pre = False
+                if self.backquote_code_style:
+                    self.out("\n" + self.pre_indent + "```")
                 if self.mark_code:
                     self.out("\n[/code]")
             self.p()
@@ -786,17 +794,23 @@ def o(
                 bq += " "
 
             if self.pre:
-                if not self.list:
+                if self.list:
+                    bq += self.list_code_indent
+
+                if not self.backquote_code_style:
                     bq += "    "
-                # else: list content is already partially indented
-                bq += "    " * len(self.list)
+
                 data = data.replace("\n", "\n" + bq)
+                self.pre_indent = bq
 
             if self.startpre:
                 self.startpre = False
-                if self.list:
+                if self.backquote_code_style:
+                    self.out("\n" + self.pre_indent + "```")
+                    self.p_p = 0
+                elif self.list:
                     # use existing initial indentation
-                    data = data.lstrip("\n")
+                    data = data.lstrip("\n" + self.pre_indent)
 
             if self.start:
                 self.space = False
@@ -952,8 +966,15 @@ def optwrap(self, text: str) -> str:
         # because of the presence of a link in it
         if not self.wrap_links:
             self.inline_links = False
+        start_code = False
         for para in text.split("\n"):
-            if len(para) > 0:
+            # If the text is between tri-backquote pairs, it's a code block;
+            # don't wrap
+            if self.backquote_code_style and para.lstrip().startswith("```"):
+                start_code = not start_code
+            if start_code:
+                result += para + "\n"
+            elif len(para) > 0:
                 if not skipwrap(
                     para, self.wrap_links, self.wrap_list_items, self.wrap_tables
                 ):

html2text/cli.py

@@ -238,6 +238,13 @@ class bcolors:
         default=config.MARK_CODE,
         help="Mark program code blocks with [code]...[/code]",
     )
+    p.add_argument(
+        "--backquote-code-style",
+        action="store_true",
+        dest="backquote_code_style",
+        default=config.BACKQUOTE_CODE_STYLE,
+        help="Multi line code block using tri-backquote style",
+    )
     p.add_argument(
         "--decode-errors",
         dest="decode_errors",
@@ -318,6 +325,7 @@ class bcolors:
     h.skip_internal_links = args.skip_internal_links
     h.links_each_paragraph = args.links_each_paragraph
     h.mark_code = args.mark_code
+    h.backquote_code_style = args.backquote_code_style
     h.wrap_links = args.wrap_links
     h.wrap_list_items = args.wrap_list_items
     h.wrap_tables = args.wrap_tables

html2text/config.py

@@ -48,6 +48,7 @@
 IMAGES_WITH_SIZE = False
 IGNORE_EMPHASIS = False
 MARK_CODE = False
+BACKQUOTE_CODE_STYLE = False
 DECODE_ERRORS = "strict"
 DEFAULT_IMAGE_ALT = ""
 PAD_TABLES = False

test/backquote_code_style.html

@@ -0,0 +1,31 @@
+<p><pre>
+def func(x):
+  if x &lt; 1:
+    return 'a'
+  return 'b'
+</pre></p>
+
+<ul>
+  <li>unordered</li>
+  <li>...</li>
+  <ol>
+    <li>ordered</li>
+    <li>code:
+      <pre>a
+b
+c</pre>
+    </li>
+    <li>...</li>
+    <ol>
+      <li>ordered</li>
+      <li>code:
+        <pre>d
+e
+f</pre>
+      </li>
+      <li>...</li>
+    </ol>
+    <li>end</li>
+  </ol>
+  <li>end</li>
+</ul>

test/backquote_code_style.md

@@ -0,0 +1,32 @@
+
+```
+def func(x):
+  if x < 1:
+    return 'a'
+  return 'b'
+
+```
+
+  * unordered
+  * ...
+    1. ordered
+    2. code: 
+       ```
+       a
+       b
+       c
+       ```
+
+    3. ...
+       1. ordered
+       2. code: 
+          ```
+          d
+          e
+          f
+          ```
+
+       3. ...
+    4. end
+  * end
+

test/mixed_nested_lists.html

@@ -18,3 +18,28 @@
   <li>end</li>
 </ul>
 
+
+<ul>
+  <li>unordered</li>
+  <li>...</li>
+  <ol>
+    <li>ordered</li>
+    <li>code:
+      <pre>a
+b
+c</pre>
+    </li>
+    <li>...</li>
+    <ol>
+      <li>ordered</li>
+      <li>code:
+        <pre>d
+e
+f</pre>
+      </li>
+      <li>...</li>
+    </ol>
+    <li>end</li>
+  </ol>
+  <li>end</li>
+</ul>

test/mixed_nested_lists.md

@@ -9,3 +9,25 @@
     1. ordered
     2. ...
   * end
+
+  * unordered
+  * ...
+    1. ordered
+    2. code: 
+           
+           a
+           b
+           c
+
+    3. ...
+       1. ordered
+       2. code: 
+              
+              d
+              e
+              f
+
+       3. ...
+    4. end
+  * end
+

test/normal.md

@@ -12,8 +12,8 @@ text to separate lists
 
   1. now with numbers
   2. the prisoner
-    1. not an _italic number_
-    2. a **bold human** being
+     1. not an _italic number_
+     2. a **bold human** being
   3. end
 
 **bold**  

test/normal_escape_snob.md

@@ -12,8 +12,8 @@ text to separate lists
 
   1. now with numbers
   2. the prisoner
-    1. not an _italic number_
-    2. a **bold human** being
+     1. not an _italic number_
+     2. a **bold human** being
   3. end
 
 **bold**  

test/preformatted_in_list.md

@@ -1,5 +1,5 @@
   * Run this command: 
-    
+        
         ls -l *.html
 
   * ?